CN102207040A

CN102207040A - Abnormal combustion detection method for spark-ignition engine, and spark-ignition engine

Info

Publication number: CN102207040A
Application number: CN2011100845894A
Authority: CN
Inventors: 松尾直也; 志志目宏二; 早川元雄; 森本博贵; 大桥美贵典
Original assignee: Mazda Motor Corp
Current assignee: Mazda Motor Corp
Priority date: 2010-03-31
Filing date: 2011-03-24
Publication date: 2011-10-05
Anticipated expiration: 2031-03-24
Also published as: DE102011012722B4; JP5565364B2; CN102207040B; JP2011226473A; DE102011012722A1; US8639432B2; US20110246049A1

Abstract

When a maximum value of vibration intensity (maximum vibration intensity) (Vmax) acquired from a vibration sensor (33) in a low engine speed/high engine load (operating region (R)) is equal to or greater than a given threshold value (X), a spark timing of a spark plug (16) is shifted from a point set in a normal state on a retard side with respect to a compression top dead center, farther toward the retard side. Then, when a maximum vibration intensity (Vmax2) acquired after the spark timing retard is greater than a maximum vibration intensity (Vmax1) acquired before the spark timing retard, it is determined that preignition occurs. An in-cylinder pressure sensor for detecting an in-cylinder pressure of an engine may be used to determine the presence or absence of the preignition, in the same manner. This technique makes it possible to reliably detect preignition using the vibration sensor, while distinguishing the preignition from knocking.

Description

Abnormal combustion detection method for spark-ignition engine, and spark-ignition engine

Technical Field

The present invention relates to a method of detecting abnormal combustion in a spark ignition engine which is provided with a vibration sensor for detecting vibration of the engine or an in-cylinder pressure sensor for detecting in-cylinder pressure and in which the ignition timing of an ignition plug is set on the retard side of compression top dead center in a normal state where abnormal combustion does not occur in a low rotation speed and high load region of the engine, and a spark ignition engine using the method.

Background

Conventionally, as shown in, for example, japanese patent laid-open publication No. 2006-46140 (hereinafter referred to as "patent document 1"), a spark ignition type engine provided with a spark plug is provided with a sensor (ion current sensor) for detecting an ion current generated by combustion of a mixed gas in a combustion chamber, and post-ignition (post-ignition) and pre-ignition (pre-ignition) that may occur during operation of the engine are detected based on a detection value of the ion current sensor.

According to patent document 1, the post-ignition is a phenomenon in which the mixture is ignited with almost no delay with respect to the ignition timing of the ignition plug. That is, in the case of normal combustion, the mixture is ignited after a predetermined delay time (ignition delay) has elapsed from the ignition timing, and when the post-combustion occurs, the mixture is self-ignited at a timing at which the ignition delay time is almost absent. On the other hand, the preignition refers to a phenomenon in which the mixed gas spontaneously ignites before the ignition timing, and the spontaneous ignition timing is earlier than the post-ignition timing.

That is, in patent document 1, if the timing at which the mixed gas is self-ignited is after the ignition timing, it is called after-ignition, and if it is before the ignition timing, it is called pre-ignition. Afterburning is lighter than pre-ignition in the sense of abnormal combustion, but may be referred to as a precursor phenomenon that will develop into pre-ignition. That is, when the afterburning occurs, the self-ignition timing is advanced, and the risk of developing pre-ignition is extremely high.

Further, when the pre-ignition is developed, severe noise and vibration occur, and if this phenomenon is continued for a long time, there is a risk of damage to the piston and the like. Thus, the pre-ignition will become a significant abnormal combustion that is not negligible. Therefore, it is desirable to be able to detect an abnormality at an early stage as much as possible to prevent pre-ignition.

Therefore, in patent document 1, it is first determined whether or not afterburning has occurred using an ion current sensor. Specifically, the peak timing of the combustion ion current is determined based on the detection value of the ion current sensor, and the presence or absence of the occurrence of the afterburning is determined based on whether or not the peak timing is advanced by a predetermined amount or more from the reference timing. If it is confirmed that the afterburning has occurred, prescribed control (e.g., increase in fuel injection amount, etc.) for suppressing this afterburning is executed even before the progress of the preignition.

In patent document 1, the self-ignition after ignition is regarded as afterburning, and the self-ignition before ignition is regarded as pre-ignition, but both cases are a phenomenon in which the self-ignition of the air-fuel mixture occurs at a timing too early with respect to a normal combustion start timing (a timing after a predetermined delay time has elapsed from the ignition timing). Therefore, in the present specification, the post-ignition and the pre-ignition are not distinguished from each other, but are referred to as the pre-ignition hereinafter.

However, when the preignition is detected using the ion current sensor as in the configuration of patent document 1, there is a possibility that the preignition cannot be detected with high accuracy in some cases. For example, patent document 1 discloses a spark plug used as an ion current sensor, but in this case, since there are cases where the ion current cannot be detected during the discharge period and the influence of LC resonance noise in the circuit, etc., there is a case where the combustion ion current cannot be detected smoothly and the preignition may be missed.

On the other hand, japanese patent laid-open publication No. 2002-339780 (hereinafter referred to as "patent document 2") discloses a technique for detecting preignition using a vibration sensor that detects vibration of an engine. Specifically, in this document, the vibration intensity and the vibration generation timing of the vibration generated in the engine main body are determined using the above-described vibration sensor, and it is determined that pre-ignition occurs when the vibration intensity exceeds a specified permissible value and the vibration generation timing is located on the advance angle side of the ignition timing.

As in patent document 2, when the preignition is detected by using the vibration sensor, the above-described problem (e.g., the detection during the discharge period is limited) that may occur when the preignition is detected by using the ion current sensor can be solved. Further, the vibration sensor has been widely used for detecting knocking, and is therefore advantageous in terms of cost.

However, as in patent document 2, if the vibration intensity and the vibration generation timing are determined using only the vibration sensor, there is a problem that if severe pre-ignition occurs, such as large vibration occurring before the ignition timing, it is not possible to detect the vibration intensity and the vibration generation timing. That is, if vibration occurs after the ignition timing, there is a possibility that the vibration is not caused by preignition of the air-fuel mixture but knocking (a phenomenon in which unburned air-fuel mixture is ignited by itself after the start of combustion), and therefore, in order to clearly judge the occurrence of preignition, it is necessary to wait until the occurrence of serious preignition such as vibration occurs before the ignition timing, which is not preferable in terms of reliability, durability, and the like of the engine.

As a method similar to the above-described method for detecting the preignition using the vibration sensor, a cylinder pressure sensor for detecting the cylinder pressure of the engine may be used. That is, when a large in-cylinder pressure exceeding the allowable value is detected and the detected period is too early, it can be determined that pre-ignition has occurred. However, even in this case, there is the same problem that if the degree of pre-ignition is not sufficiently severe, detection cannot be distinguished from knocking.

Disclosure of Invention

The present invention has been made in view of the above circumstances, and an object thereof is to provide a method and an engine using the method, in which: the vibration sensor or the in-cylinder pressure sensor is used to distinguish the preignition and the knocking which are likely to occur during the engine operation and to appropriately detect the preignition.

In order to achieve the above object, a method of the present invention is a method of detecting abnormal combustion in a spark ignition engine provided with a vibration sensor that detects vibration of the engine or an in-cylinder pressure sensor that detects an in-cylinder pressure, the engine setting an ignition timing of a spark plug at a normal time in a low rotation speed and high load region and in which abnormal combustion does not occur on a retard angle side of a compression top dead center, the method including the steps of: determining whether or not a maximum value of the vibration intensity obtained from the vibration sensor or a maximum value of the in-cylinder pressure obtained from the in-cylinder pressure sensor in a low rotation speed and high load region is equal to or greater than a predetermined threshold value; changing the ignition timing of the spark plug from the ignition timing of the normal time on the retard side of the compression top dead center to a further retard side when the maximum value of the vibration intensity or the maximum value of the in-cylinder pressure is equal to or greater than the predetermined threshold value; when the maximum value of the vibration intensity obtained from the vibration sensor after the ignition timing is changed to the retard side or the maximum value of the in-cylinder pressure obtained from the in-cylinder pressure sensor is set to the maximum detection value after the ignition delay and the maximum value of the vibration intensity obtained before the ignition timing is changed to the retard side or the maximum value of the in-cylinder pressure is set to the maximum detection value before the ignition delay, it is determined that preignition of the mixture is caused if the maximum detection value after the ignition delay is larger than the maximum detection value before the ignition delay.

An engine according to the present invention is a spark ignition type engine including a vibration sensor for detecting vibration of the engine or an in-cylinder pressure sensor for detecting an in-cylinder pressure, the engine setting an ignition timing of a spark plug at a retard angle side of a compression top dead center in a normal time in a low rotation speed and high load region and in which abnormal combustion does not occur, the engine including: a control unit that controls a timing of spark discharge of the spark plug and receives information of a vibration intensity detected by the vibration sensor or information of an in-cylinder pressure detected by the in-cylinder pressure sensor; wherein the control means executes control to change the ignition timing of the spark plug from the ignition timing of the normal time on the retard angle side located at a compression top dead center to the retard angle side further when a maximum value of the vibration intensity acquired from the vibration sensor or a maximum value of the in-cylinder pressure acquired from the in-cylinder pressure sensor in a low rotation speed and high load region is equal to or more than a predetermined threshold value, and if the maximum value of the vibration intensity acquired from the vibration sensor after the ignition timing is changed to the retard angle side or the maximum value of the in-cylinder pressure acquired before the ignition timing is changed to the retard angle side is set as a maximum detection value after an ignition delay and if the maximum detection value after the ignition delay is larger than the maximum detection value before the ignition delay when the maximum value of the vibration intensity acquired before the ignition delay or the maximum value of the in-cylinder pressure acquired before the ignition delay side is set as a maximum detection value before the ignition delay, it is determined that preignition in which the mixture is ignited prematurely occurs.

According to the method and the engine using the method, the preignition that may occur during the operation of the engine can be distinguished from the knocking by the vibration sensor or the in-cylinder pressure sensor, and the preignition can be appropriately detected.

Drawings

Fig. 1 is a diagram showing an overall configuration of a spark ignition engine according to an embodiment of the present invention.

Fig. 2 is a block diagram showing the control system of the engine.

Fig. 3 is a diagram showing a specific operation region in which preignition is likely to occur.

Fig. 4 is a diagram showing the distribution (temporal change) of the amount of heat generated at the time of pre-ignition and at the time of normal combustion.

Fig. 5 is a graph showing a comparison between a change in the in-cylinder pressure when the pre-ignition occurs and a change in the in-cylinder pressure when the knocking occurs.

Fig. 6 is a diagram illustrating a waveform input from the vibration sensor when pre-ignition occurs.

Fig. 7 is a waveform illustrating an input from the vibration sensor when knocking occurs.

Fig. 8 is a graph showing how the magnitude of the maximum vibration intensity and the measured timing thereof change when the ignition timing is retarded at the time of preignition and occurrence of knocking.

Fig. 9 is a flowchart showing the contents of the control operation for detecting preignition and knocking.

Fig. 10 is a flowchart showing the contents of the control operation executed after the detection result of fig. 9 is obtained.

Fig. 11 is a subroutine showing specific contents of the preignition avoidance control included in the flowchart of fig. 10.

Fig. 12 is a subroutine showing the details of the recovery control included in the flowchart of fig. 10.

Fig. 13 is a diagram showing the relationship between the retard angle amount of the closing timing of the intake valve and the effective compression ratio.

Fig. 14 is a diagram showing how the retard angle amount of the closing timing of the intake valve required to lower the effective compression ratio by 0.5 changes in accordance with the value of the present retard angle amount.

Fig. 15 (a) and (b) show injection timings of fuel, fig. 15 (a) shows an injection timing in a normal state, and fig. 15 (b) shows an injection timing in a pre-ignition state.

Fig. 16 is a time chart showing an operation example of the preignition avoidance control in time series.

Fig. 17 is a time chart showing an operation example of the recovery control in time series.

Fig. 18 (a) to (c) are views for explaining modifications of the present invention.

Detailed Description

(1) Integral structure of engine

Fig. 1 is a diagram showing an overall configuration of an engine according to an embodiment of the present invention. The engine shown in the figure is a reciprocating piston type multi-cylinder gasoline engine, and is mounted on a vehicle as a power source for driving. The engine body 1 of this engine includes: a cylinder block 3 having a plurality of cylinders 2 (only 1 of which is shown in the figure) arrayed in a direction orthogonal to the paper surface; a cylinder head 4 provided on an upper surface of the cylinder block 3; and a piston 5 inserted into each cylinder 2 to be slidable back and forth. The fuel supplied to the engine body 1 may be a fuel containing gasoline as a main component, and may be entirely gasoline itself or a fuel containing ethanol (ethyl alcohol) or the like in gasoline.

The piston 5 is connected to a crankshaft 7 via a connecting rod 8, and the crankshaft 7 rotates about a central axis in response to the reciprocation of the piston 5.

A combustion chamber 6 is formed above the piston 5, an intake port 9 and an exhaust port 10 are opened in the combustion chamber 6, and an intake valve 11 and an exhaust valve 12 for opening and closing the

ports

9 and 10 are provided in the cylinder head 4, respectively. The intake valve 11 and the exhaust valve 12 are driven to open and close in conjunction with rotation of the crankshaft 7 by

valve operating mechanisms

13 and 14 including a pair of camshafts (not shown) and the like disposed in the cylinder head 4.

A VVT15 is incorporated in the valve train 13 for the intake valve 11. The VVT15 is called a Variable Valve Timing Mechanism (Variable Valve Timing Mechanism) and is a Variable Mechanism for variably setting the operation Timing of the intake Valve 11.

Various types of VVTs have been put to practical use as the VVT15, and a hydraulic variable mechanism may be used as the VVT 15. Although not shown, the hydraulic variable mechanism includes a driven shaft coaxially disposed on a camshaft for the intake valve 11, and a plurality of liquid chambers disposed in a circumferential direction between the camshaft and the driven shaft, and a predetermined pressure difference is formed between the liquid chambers, thereby forming a phase difference between the camshaft and the driven shaft. Then, by variably setting this phase difference within a specified angle range, the operation timing of the intake valve 11 is continuously changed.

Further, as the VVT15, a variable mechanism of a type that changes the closing timing of the intake valve 11 by changing the valve lift amount may be provided. Further, such a lift type variable mechanism and the above-described phase type variable mechanism may be used in combination.

The cylinder head 4 of the engine body 1 is provided with 1 group of ignition plugs 16 and injectors 18 for each cylinder 2.

The injector 18 is provided so as to face the combustion chamber 6 from the side surface on the intake side, and injects fuel (gasoline) supplied from a fuel supply pipe outside the drawing from the distal end portion. In an intake stroke of the engine or the like, fuel is injected from the injector 18 into the combustion chamber 6, and the injected fuel is mixed with air, whereby a mixture gas of a desired air-fuel ratio is generated in the combustion chamber 6.

The ignition plug 16 is provided so as to face the combustion chamber 6 from above, and emits a spark from a distal end portion in response to power supply from an ignition circuit outside the figure. At a predetermined timing set in the vicinity of the compression top dead center, a spark is released from the spark plug 16 to start combustion of the air-fuel mixture.

The cylinder block 3 is provided with an engine speed sensor 30 that detects the rotational speed of the crankshaft 7 as the rotational speed of the engine.

The cylinder block 3 is provided with a vibration sensor 33 for detecting vibration of the cylinder block 3. The detection value of the vibration sensor 33 is used to detect abnormal combustion generated in the engine.

Specifically, in the present embodiment, two types of abnormal combustion, i.e., knocking and preignition, are detected based on the detection value of the vibration sensor 33. Here, knocking means a phenomenon in which, after combustion of the air-fuel mixture is started by spark ignition, an unburned component (residual exhaust gas) of the air-fuel mixture is self-ignited during propagation of a flame. On the other hand, the term "pre-ignition" refers to a phenomenon in which the mixed gas is self-ignited before a normal combustion start timing by spark ignition (i.e., regardless of spark ignition). When knocking or preignition occurs, relatively large vibration occurs in the cylinder block 3 due to rapid combustion pressure fluctuation or the like, and therefore, in the present embodiment, such vibration of the cylinder block 3 is examined based on the detection value of the vibration sensor 33, and knocking or preignition is detected.

An ion current sensor 34 for detecting a flame generated by combustion of the gas mixture in the combustion chamber 6 is provided in the vicinity of the ignition plug 16. The ion current sensor 34 has an electrode to which a bias voltage of, for example, about 100V is applied, and detects a flame by detecting an ion current generated when a flame is formed around the electrode.

By detecting the flame using the ion current sensor 34, the occurrence of preignition can be detected in the same manner as the vibration sensor 33. That is, when the air-fuel mixture is forcibly combusted by the spark ignition, if the combustion state is normal, the combustion should be started after a predetermined delay time has elapsed from the spark ignition, but if the pre-ignition occurs, the air-fuel mixture is self-ignited early regardless of the spark ignition, and therefore the combustion is started before the above-mentioned normal combustion start timing (the timing after the predetermined delay time has elapsed from the spark ignition). Therefore, if the flame is detected by the ion current sensor 34 and the detected time (the flame generation time) is earlier than the normal combustion start time, it is determined that preignition has occurred. As described above, in the present embodiment, two types of sensors, i.e., the ion current sensor 34 and the vibration sensor 33, are provided as sensors for detecting preignition, and the use of these two types of sensors makes it possible to more reliably detect preignition.

However, only preignition that can be detected by using the ion current sensor 34 described above is detected, and knocking cannot be detected. That is, as described above, knocking is a phenomenon in which, after flame is once generated from spark ignition as a start, an unburned component (residual exhaust gas) of the mixed gas spontaneously ignites during its propagation, and therefore, even if knocking occurs, the occurrence timing of the flame is not determined as if it were the case in the conventional case, even if the occurrence timing of the flame is examined by the ion current sensor 34. Therefore, in detecting knocking, only the detection value of vibration sensor 33 is used, and ion current sensor 34 is not used.

An intake passage 20 and an exhaust passage 21 are connected to the intake port 9 and the exhaust port 10 of the engine body 1, respectively. That is, intake air (fresh air) from the outside is supplied to the combustion chamber 6 through the intake passage 20, and burned gas (exhaust gas) generated in the combustion chamber 6 is discharged to the outside through the exhaust passage 21.

The intake passage 20 is provided with a throttle valve 22 for adjusting the flow rate of intake air flowing into the engine body 1 and an airflow sensor 31 for detecting the flow rate of intake air.

The throttle valve 22 is an electronically controlled throttle valve, and is electrically opened and closed in accordance with the degree of opening of an accelerator pedal out of the drawing that is depressed by the driver. That is, an accelerator opening sensor 32 (fig. 2) is provided in the accelerator pedal, and an electronic actuator outside the figure drives the throttle valve 22 to open and close according to the opening degree of the accelerator pedal (accelerator opening degree) detected by the accelerator opening sensor 32.

An exhaust gas purifying catalytic converter 23 is provided in the exhaust passage 21. The catalytic converter 23 incorporates, for example, a three-way catalyst, and harmful components in the exhaust gas passing through the exhaust passage 21 are purified by the three-way catalyst.

(2) Control system

Fig. 2 is a block diagram showing a control system of the engine. The ECU40 shown in this drawing is a control unit for comprehensively controlling each part of the engine, and includes a well-known CPU, ROM, RAM, and the like.

Detection signals from various sensors are input to the ECU 40. That is, the ECU40 is electrically connected to the engine speed sensor 30, the airflow sensor 31, the accelerator opening sensor 32, the vibration sensor 33, and the ion current sensor 34, and sequentially inputs information such as the engine speed Ne, the intake air amount Qa, the accelerator opening AC, the vibration intensity (acceleration) Va, and the ion current value Io to the ECU40 as detection values of the sensors 30 to 34.

The ECU40 is also electrically connected to the VVT15, the ignition plug 16, the injector 18, and the throttle valve 22, and outputs drive control signals to these devices.

Next, more specific functions of the ECU40 will be described, and the ECU40 includes the storage means 41, the abnormal combustion determination means 42, the ignition control means 43, the fuel control means 44, and the VVT control means 45 as its main components.

The storage unit 41 stores various data or programs required for controlling the engine. As an example thereof, the range of the specific operation region R shown in fig. 3 is stored in the storage unit 41. The specific operating range R is an operating range in which pre-ignition is likely to occur, and is set near the maximum load line WOT (i.e., high load) and closer to the low rotation speed.

That is, as described above, since the pre-ignition is a phenomenon in which the air-fuel mixture is self-ignited before the normal combustion start timing by the spark ignition, the pre-ignition is most likely to occur in a low-revolution high-load region in which the air in the combustion chamber 6 is increased in temperature and pressure and the period during which the fuel receives heat from the air is increased. Therefore, as shown in fig. 3, a region in which the engine speed Ne is relatively low and the load Ce is high is set as the specific operation region R in which preignition is likely to occur.

The abnormal combustion determination means 42 determines the presence or absence of preignition or knocking based on the detection values of the vibration sensor 33 and the ion current sensor 34. Specifically, when the operating state of the engine is in the specific operating region R, the abnormal combustion determination means 42 determines the occurrence timing of the flame based on the detection value (ion current value Io) of the ion current sensor 34, and compares the determination timing with the normal combustion start timing to determine whether or not the preignition has occurred. The abnormal combustion determination means 42 determines whether preignition or knocking has occurred by examining the maximum value of the vibration intensity and the occurrence timing thereof based on the detection value (vibration intensity Va) of the vibration sensor 33 (see item (3) described later in detail).

The ignition control means 43 outputs a power supply signal to an ignition circuit of the ignition plug 16 at a predetermined timing set in advance in accordance with an operating state of the engine, thereby controlling a timing (ignition timing) at which the ignition plug 16 performs spark discharge and the like.

For example, in the specific operating region R in the low rotation speed/high load region of the engine, the spark plug 16 is controlled to perform spark ignition at a timing slightly later than the compression top dead center. However, if vibration of a predetermined level or more is input from the vibration sensor 33 in the specific operation region R, the ignition control means 43 displaces the ignition timing further to the retard side than the timing (timing slightly later than the compression top dead center). This is to determine whether the vibration of a predetermined level or higher inputted from the vibration sensor 33 is caused by pre-ignition or knocking.

That is, although the retarded angle of the ignition timing (change to the retarded angle side) acts in the direction of suppressing the knock, the ignition control means 43 does not particularly affect the pre-ignition (the reason thereof will be described later), and when the vibration of a predetermined level or more occurs, the ignition control means intentionally retards the ignition timing in order to determine whether the cause is the pre-ignition or the knock. Then, the abnormal combustion determination means 42 examines the change in the vibration after the ignition timing is retarded, and determines whether preignition or knocking has occurred based on the result.

The fuel control means 44 controls the injection amount and the injection timing of the fuel injected from the injector 18 into the combustion chamber 6. More specifically, the fuel control unit 44 calculates a target fuel injection amount and injection timing based on information such as the engine rotation speed Ne input from the engine rotation speed sensor 30 and the intake air amount Qa input from the airflow sensor 31, and controls the valve opening timing and the valve opening period of the injector 18 based on the calculation result.

In particular, when pre-ignition is detected in the specific operation region R, the fuel control means 44 increases the fuel injection amount injected from the injector 18 to perform control for enriching the air-fuel ratio in the cylinder. Such control is performed in order to suppress the occurrence of pre-ignition by reducing the in-cylinder temperature by injecting a relatively large amount of fuel. Further, the above-described fuel control unit 44 executes control of retarding a part of the fuel, which should originally be injected in the intake stroke, to the late stage of the compression stroke, if necessary (i.e., injecting the fuel in the intake stroke and the compression stroke in a divided manner). Thereby, the in-cylinder temperature particularly near the compression top dead center is reduced, and the period of heating of the fuel is shortened, so that an environment in which pre-ignition is more difficult to occur can be created.

The VVT control means 45 variably sets the effective compression ratio of the engine by driving the VVT15 to change the closing timing of the intake valve 11. That is, the closing timing of the intake valve 11 is normally set in the vicinity of the retarded side of intake bottom dead center (timing slightly beyond intake bottom dead center), and by setting such timing, the sucked air is hardly blown back to the intake port 9, and the substantial compression ratio (effective compression ratio) of the engine can be brought close to the geometric compression ratio. In contrast, if the closing timing of the intake valve 11 is set to be significantly delayed from intake bottom dead center, the effective compression ratio of the engine is lowered accordingly, causing a considerable amount of backflow of intake air. The VVT control means 45 drives the VVT15 to increase or decrease the delay amount (delay angle amount) of the closing timing of the intake valve 11, thereby variably setting the effective compression ratio of the engine.

In particular, the VVT control means 45 executes control for retarding the closing timing of the intake valve 11 and lowering the effective compression ratio as necessary when pre-ignition is detected in the specific operation region R. This mainly reduces the in-cylinder pressure (pressure in the combustion chamber 6) and suppresses pre-ignition.

Note that the "closing timing of the intake valve 11" mentioned in the above description means a closing timing in the case where a section other than the slope portion of the lift curve (the buffer section in which the lift amount gradually increases) is defined as the open period of the valve, and not a timing at which the lift amount of the intake valve 11 is completely zero.

(3) Method for determining preignition and knocking

Next, more specific steps when the abnormal combustion determination means 42 determines the occurrence of the preignition and the knocking will be described.

First, a procedure for detecting the preignition time by using the ion current sensor 34 will be described. Fig. 4 is a diagram showing the distribution (temporal change) of the amount of heat generated at the time of pre-ignition and at the time of normal combustion. In this figure, IG represents spark ignition, and the amount of heat generation during normal combustion that starts with this spark ignition IG is a waveform J0 of a solid line. As described above, the timing of the spark ignition IG is set to a timing slightly later than the compression top dead center in the specific operating region R where the pre-ignition is likely to occur. Therefore, the IG position in the figure is set on the delay angle side of the compression Top Dead Center (TDC), and in the figure, it is about 5 ° CA beyond the compression top dead center (ATDC).

In the waveform J0 during normal combustion that is initiated by the spark ignition IG, when the state (substantial combustion start timing) in which combustion has progressed to such an extent that the ion current sensor 34 can detect flame is t0, this time t0 is later than the time of spark ignition IG by a predetermined crank angle. That is, during normal combustion, the flame kernel generated by spark ignition gradually expands from the center toward the periphery, so the substantial combustion start timing t0 is delayed to some extent from spark ignition IG.

On the other hand, the distribution of the amount of heat generation when pre-ignition occurs is shown by the dashed-dotted line waveforms J1 to J3. The waveform J1 indicates light preignition, the waveform J2 indicates moderate preignition, and the waveform J3 indicates heavy preignition, and if the substantial combustion start timings in each case are t1, t2, and t3, respectively, the crank angle position thereof is shifted to the advance side from the normal combustion start timing t 0. That is, when the pre-ignition occurs, the mixture is self-ignited, so that the combustion cannot be controlled by the spark ignition, and the combustion is started before the normal combustion start time t 0. Further, as the start of combustion becomes earlier, the combustion becomes more rapid, and the combustion period becomes shorter.

Moreover, the pre-ignition has the property that it progresses from a mild pre-ignition (J1) to a severe pre-ignition (J3) if left alone. That is, if the pre-ignition occurs, the temperature of the combustion chamber 6 is accelerated to create an environment in which the self-ignition is more likely to occur, and therefore the pre-ignition progresses in a chain. In particular, when severe pre-ignition (J3) occurs, the combustion becomes extremely rapid, the engine generates extreme noise and vibration, and the piston and the like are damaged.

Therefore, it is necessary to appropriately detect the occurrence of pre-ignition and take necessary measures (e.g., enrichment of the air-fuel ratio, etc.) at least before the development of severe pre-ignition as described above. Therefore, in the present embodiment, as one means for detecting the preignition, the flame is detected by using the ion current sensor 34, and the presence or absence of the preignition is determined based on the detected timing (the timing of the occurrence of the flame). Specifically, when the ion current sensor 34 detects a flame at a stage earlier than the normal combustion start time t0 by a predetermined time or more, this is detected as preignition. In this case, it is preferable that the flame detection timing of the ion current sensor 34 is advanced to, for example, about t1, i.e., it is determined as preignition, in order to detect preignition at a stage as light as possible.

However, as described above, the occurrence of pre-ignition is detected not only by the ion current sensor 34 but also by the vibration sensor 33. In the present embodiment, the vibration sensor 33 is used not only to detect preignition but also to detect knocking. Next, a detection procedure using this vibration sensor 33 will be described.

Fig. 5 is a graph showing a comparison between a change in the in-cylinder pressure when the pre-ignition occurs and a change in the in-cylinder pressure when the knocking occurs. In fig. 5, the change in the cylinder pressure at the time of occurrence of pre-ignition is represented by a waveform Pp, and the change in the cylinder pressure at the time of occurrence of knocking is represented by a waveform Pn. In this figure, in order to clearly show the difference between the two waveforms Pp and Pn, the waveform Pp at the time of occurrence of the pre-ignition is exemplified by the change in the in-cylinder pressure when the pre-ignition has progressed to a certain degree (when the pre-ignition has progressed to a severe degree or a degree close to a severe degree).

As is clear from the observation of the above waveform Pp, when the pre-ignition has progressed to a certain extent, the in-cylinder pressure rises greatly in the vicinity of the compression top dead center, and the rising pressure converges in a relatively short period. On the other hand, when knocking occurs, as shown by the waveform Pn, a peak of the waveform in which the in-cylinder pressure rises sharply occurs at a position greatly shifted to the retarded angle side from the position at the time of preignition. That is, since knocking is a phenomenon in which the remaining unburned gas mixture (remaining exhaust gas) is self-ignited when combustion progresses to a certain extent, a sharp rise in pressure due to the self-ignition occurs at the final stage of the combustion process, and the peak of the waveform is shifted further toward the retarded angle side.

Fig. 6 and 7 show how the vibration waveform is input from the vibration sensor 33 when the change in the in-cylinder pressure shown in fig. 5 occurs due to the occurrence of pre-ignition and knocking. Here, the vibration waveform represents a change in the vibration intensity Va according to the crank angle CA, using the vibration intensity (acceleration) Va input from the vibration sensor 33 as the vertical axis and the crank angle CA as the horizontal axis.

If the waveforms of fig. 6 and 7 are compared, it is understood that the maximum value Vmax of the detected vibration intensity Va (hereinafter, simply referred to as the maximum vibration intensity Vmax) is larger in the vibration waveform at the time of occurrence of pre-ignition (fig. 6) than in the vibration waveform at the time of occurrence of knocking (fig. 7), and the measurement timing thereof is earlier. The reason for this is considered that, in the case of pre-ignition that has progressed to a certain degree as shown in fig. 5, the portion where the change in the in-cylinder pressure is most severe (i.e., the peak portion of the waveform) has a larger fluctuation width than the case of knocking, and it occurs rather close to the advance angle side.

It is thus known that the maximum detected vibration intensity Vmax and its measured period are relatively well characterized when pre-ignition has progressed to a certain extent. However, if the preignition is not so severe (for example, if a slight preignition occurs as in the waveform J1 shown in fig. 4), the maximum vibration intensity Vmax and the measured time thereof are not greatly different from the case of knocking, and there is a risk that the preignition and knocking cannot be detected clearly by simply comparing the waveforms of the vibration intensity Va.

Therefore, in the present embodiment, the vibration sensor 33 detects the maximum vibration intensity Vmax equal to or greater than a predetermined threshold value, intentionally retards the ignition timing to discriminate between the occurrence of pre-ignition and the occurrence of knocking when pre-ignition or knocking is suspected, and discriminates whether pre-ignition or knocking is performed based on a subsequent change in the maximum vibration intensity Vmax.

That is, in the specific operating region R which is close to the low rotation speed and has a high load and in which the pre-ignition is likely to occur, the ignition timing is set to a timing slightly later than the compression top dead center (for example, about 5 ° ATDC) in the normal state, but when the maximum vibration intensity Vmax equal to or higher than a predetermined threshold value is detected by the vibration sensor 33, the ignition timing is retarded by a predetermined amount with respect to the timing, and the spark ignition is performed at a timing further retarded with respect to the compression top dead center. Next, the abnormal combustion determination means 42 examines how the maximum vibration intensity Vmax changes according to the retarded ignition timing, and determines the occurrence of pre-ignition or knocking.

For example, if knocking occurs, the combustion starts on the more retarded side of compression top dead center (i.e., in a state where the in-cylinder temperature/pressure is more reduced) by retarding the ignition timing as described above, and spontaneous ignition of the unburned gas mixture (residual exhaust gas) is less likely to occur in the subsequent combustion process. Therefore, if the ignition timing is retarded during occurrence of knocking, the degree of knocking will be reduced and the occurrence timing thereof will become retarded. Accordingly, there occurs a phenomenon that the magnitude of the maximum vibration intensity Vmax detected by the vibration sensor 33 is decreased and the measured time is delayed.

The symbol "x" in fig. 8 indicates how the maximum vibration intensity Vmax detected by the vibration sensor 33 changes by gradually retarding the ignition timing when knocking occurs. According to this diagram, as the ignition timing is retarded, the plot (mark "x") of the maximum vibration intensity Vmax gradually moves in the downward and rightward direction. That is, as the ignition timing is retarded, the magnitude of the maximum vibration intensity Vmax gradually decreases, and the crank angle at the time when Vmax is detected gradually shifts to the retard side. The value X on the vertical axis in fig. 8 is a threshold value for determining whether or not to retard the ignition timing, and when the maximum vibration intensity Vmax equal to or greater than the threshold value X is detected, the ignition timing is retarded.

As described above, although knocking can be suppressed by retarding the ignition timing, if preignition occurs, the mixture will self-ignite regardless of the ignition timing, so even if the ignition timing is retarded, autoignition will still occur, and preignition will not be suppressed. Instead, as explained based on fig. 4, once the pre-ignition occurs, the pre-ignition gradually progresses with the passage of time, resulting in the early stage of the combustion start and the abrupt combustion. In fig. 8, the reason is that the mark "Δ" indicating the maximum vibration intensity Vmax at the time of occurrence of pre-ignition is gradually shifted in the left upper direction. That is, at the time of occurrence of pre-ignition, the magnitude of the maximum vibration intensity Vmax gradually increases with the elapse of time regardless of the delay of the ignition timing, and the measured timing gradually advances.

As is apparent from the above description, when the pre-ignition occurs, the increase in the maximum vibration intensity Vmax and the advance of the measurement timing are observed even if the ignition timing is retarded, and when the knocking occurs, the decrease in the maximum vibration intensity Vmax and the retard of the measurement timing are observed as the ignition timing is retarded. Therefore, in the present embodiment, it is determined whether preignition or knocking has occurred based on the change in maximum vibration intensity Vmax (the magnitude thereof and the change in the measured timing) associated with the late ignition timing. With this, even if the vibration sensor 33 is used, it is possible to accurately determine whether the pre-ignition or the knocking is present.

(4) Control actions

Next, the control operation of the ECU40 having the above-described functions will be described based on the flowcharts of fig. 9 to 12. Here, the description will be given mainly on the detection of preignition and knocking and the avoidance operation when they are detected.

When the processing shown in the flowchart of fig. 9 is started, first, control of reading various sensor values is executed (step S1). Specifically, the engine speed Ne, the intake air amount Qa, the accelerator opening AC, the vibration intensity Va, and the ion current value Io are read from the engine speed sensor 30, the airflow sensor 31, the accelerator opening sensor 32, the vibration sensor 33, and the ion current sensor 34, respectively, and input to the ECU 40.

Next, control is executed to determine whether or not the current engine operating state is within the specific operating region R shown in fig. 3, based on the information read in step S1 described above (step S2). Specifically, it is determined whether or not the engine speed Ne read in step S1 and the engine load Ce calculated based on the intake air amount Qa (or the accelerator opening degree AC) are both included in the range of the specific operation region R in fig. 3.

If it is determined as no in step S2 and it is confirmed that the operation is outside the specific operation region R, since pre-ignition is not likely to occur, the routine operation is maintained without the need for the processing (determination of abnormal combustion and avoidance control thereof) after step S3 (which will be described later) (step S32 in fig. 10). That is, the injection amount and injection timing of the fuel, the operation timing of the intake valve 11, and the like are controlled in accordance with normal target values set in advance in accordance with the operating state.

On the other hand, when it is determined as yes in the above-described step S2 to confirm that the operation is within the specific operation region R, control is performed to determine whether the timing of occurrence of flame is earlier than normal, that is, to determine whether preignition occurs, based on the ion current value Io read in the above-described step S1 (step S3). Specifically, it is determined that preignition has occurred when the flame generation timing determined based on the ion current value Io is earlier than a normal combustion start timing (timing slightly later than spark ignition; e.g., time t0 in fig. 4) stored in advance by a predetermined time or more.

When it is determined as "yes" in step S3 and it is confirmed that the preignition has occurred, control is performed to input "1" indicating the occurrence of the preignition to the abnormal combustion flag Fabnrm (the default value of which is 0) for recording the combustion state (step S4).

On the other hand, when it is determined as no in the above-described step S3, that is, when pre-ignition is not detected based on the ion current value Io, control is performed to acquire the maximum value (maximum vibration intensity) Vmax from the information of the vibration intensity Va read from the vibration sensor 33 in the above-described step S1 and store it as Vmax1 (step S5). Next, it is determined whether or not the stored maximum vibration intensity Vmax1 is equal to or greater than a preset threshold value X (see fig. 8) (step S6).

If it is determined as yes at step S6 and it is confirmed that the maximum vibration intensity Vmax1 is equal to or greater than the threshold value X, control is performed to retard (retard) the ignition timing of the ignition plug 16 by a predetermined amount (step S7). As described above, since the normal ignition timing in the specific operation region R is set to a timing slightly later than the compression top dead center (for example, about ATDC5 °), the retard amount from the compression top dead center to the ignition timing is further increased by the retard of the ignition timing.

When the retard angle of the ignition timing is performed as described above, control is performed in which the maximum vibration intensity Vmax is acquired from the information input from the vibration sensor 33 in the state after the retard angle, and the value is stored as Vmax2 (steps S8, S9). Next, it is determined whether the stored maximum vibration intensity Vmax2 is greater than the maximum vibration intensity Vmax1 stored in the previous step S5 (i.e., the maximum vibration intensity stored before the ignition timing is retarded) (step S10). Hereinafter, the maximum vibration intensity Vmax2 stored after the ignition timing is delayed will be referred to as "maximum vibration intensity Vmax2 after the ignition delay", and the maximum vibration intensity Vmax1 stored before the ignition timing is delayed will be referred to as "maximum vibration intensity Vmax1 before the ignition delay". At this time, the maximum vibration intensity Vmax2 after the ignition delay corresponds to the maximum detection value after the ignition delay according to the present invention, and the maximum vibration intensity Vmax1 before the ignition delay corresponds to the maximum detection value before the ignition delay according to the present invention.

If it is determined as "yes" in step S10, that is, if (maximum vibration intensity Vmax2 after ignition delay) > (maximum vibration intensity Vmax1 before ignition delay) is confirmed, since the maximum vibration intensity Vmax increases even if the ignition timing is delayed, "1" indicating that pre-ignition occurs is input to the abnormal combustion flag Fabnrm (step S4). That is, as indicated by the mark "Δ" in fig. 8, when the preignition occurs, the preignition cannot be suppressed even if the ignition timing is retarded, and the maximum vibration intensity Vmax increases, so that when Vmax2 > Vmax1 is reached as described above, it can be determined that the preignition has occurred, and as a result, the abnormal combustion mark Fabnrm is set to 1.

On the other hand, when it is determined as "no" in the above-described step S10, that is, when it is confirmed that (the maximum vibration intensity Vmax2 after the ignition delay) is ≦ the maximum vibration intensity Vmax1 before the ignition delay), control is further performed to determine whether or not the measured timing of the maximum vibration intensity Vmax2 after the ignition delay is earlier than the measured timing of the maximum vibration intensity Vmax1 before the ignition delay (step S11).

When it is determined that the determination in step S11 is yes and it is confirmed that the measurement timing of Vmax2 is earlier than the measurement timing of Vmax1, the measurement timing of the maximum vibration intensity Vmax is advanced even if the ignition timing is retarded, and therefore "1" indicating the occurrence of pre-ignition is input to the abnormal combustion flag Fabnrm (step S4), as in the case of yes in step S10. That is, as indicated by the mark "Δ" in fig. 8, when the measurement timing of the maximum vibration intensity Vmax is advanced regardless of the retarded ignition timing, it can be determined that the preignition has occurred, and therefore the abnormal combustion mark Fabnrm is set to 1.

As shown in steps S10 and S11, in the present embodiment, after the ignition timing is delayed, it is first determined whether the magnitude of the maximum vibration intensity Vmax is increased (S10), and even if it is determined that Vmax is not increased, it is further determined whether the measured timing of Vmax is advanced (S11). When it is determined as yes in both of steps S10 and S11, it is determined that preignition has occurred. That is, as shown by the symbol "Δ" in fig. 8, when the preignition occurs, the phenomenon that the maximum vibration intensity Vmax increases and the measurement timing thereof is advanced is observed, but depending on the case, only one of the phenomena may be observed, and therefore, if either one of the two determinations of steps S10 and S11 is "yes", the preignition is determined.

Next, a control operation when it is determined as no in step S11 will be described. At this time, since the maximum vibration intensity Vmax decreases as the ignition timing is retarded, and the measurement timing thereof becomes retarded, in the next step S12, control is performed in which "2" indicating that knocking has occurred is input to the abnormal combustion flag Fabnrm. That is, as shown by the mark "x" in fig. 8, when the ignition timing is retarded during the occurrence of knocking, the maximum vibration intensity Vmax will decrease, and the measured timing thereof is retarded, so when the above phenomenon is seen, it can be determined that knocking has occurred, and as a result, the abnormal combustion mark Fabnrm is made 2.

When it is determined as "no" in step S6, that is, when it is confirmed that maximum vibration intensity Vmax1 before ignition delay is smaller than threshold value X, it is considered that neither preignition nor knocking has occurred, and therefore control is performed in which "0" indicating that the combustion state is normal is input to abnormal combustion flag Fabnrm (step S13).

As described above, in the flowchart of fig. 9, when the operating state of the engine is in the specific operating region R, it is determined whether or not preignition or knocking has occurred based on the detection values of the ion current sensor 34 and the vibration sensor 33, and any of the values "0", "1", and "2" is input to the abnormal combustion flag Fabnrm according to the result of determination.

Fig. 10 shows a process following the flowchart of fig. 9 described above. When the processing shown in the present drawing is started, it is first determined whether or not the abnormal combustion flag Fabnrm is 1 (step S20), and when it is determined that "yes" (Fabnrm is 1) and it is confirmed that the preignition has occurred, as control for avoiding this phenomenon, preignition avoidance control is executed (step S21).

Next, specific contents of the preignition avoidance control in step S21 will be described with reference to fig. 11. After this preignition avoidance control is started, first, control is executed to determine whether the currently set in-cylinder air-fuel ratio (a/F) is smaller than 11 (step S40). Here, the determination threshold value (a/F ═ 11) is a threshold value that is allowed when the air-fuel ratio is made rich in step S42, which will be described later. If the air-fuel ratio is enriched to a value smaller than a/F11, smoke may be generated and fuel consumption is disadvantageous, and therefore a/F11 is set as the threshold value for enrichment.

In the specific operation region R, the in-cylinder air-fuel ratio is initially set to a value that is at the stoichiometric air-fuel ratio (═ 14.7) or slightly richer than this, and is at an air-fuel ratio that is leaner than the threshold value (═ 11). Therefore, the first determination in step S40 is yes, and the process proceeds to step S42 to perform the control of enriching the air-fuel ratio. Specifically, by increasing the injection amount of the fuel injected from the injector 18, the in-cylinder air-fuel ratio is made richer by a specified amount than the air-fuel ratio currently set.

The enrichment of the air-fuel ratio is performed in stages by a plurality of times. For example, if the current air-fuel ratio is a/F14.7 (stoichiometric air-fuel ratio), the air-fuel ratio is made rich to a smaller a/F12.5, and if this is not possible to avoid preignition, the air-fuel ratio is made rich to a smaller a/F11 (threshold). On the contrary, if the pre-ignition is avoided at the time when a/F is 12.5, the enrichment is stopped at this time.

If the pre-ignition continues after the air-fuel ratio is enriched to the threshold value a/F at step S42 is set to 11, the determination of step S40 is no, and therefore, at the next step S41, control is executed to determine whether the closing timing (IVC) of the intake valve 11 currently set is earlier than the closing timing (latest timing) Tx at which the IVC is retarded to the maximum at step S43 described later. The determination threshold value, i.e., the latest timing Tx, is set to a timing at which the effective compression ratio of the engine is lowered to a certain extent with respect to the geometric compression ratio and the reverse flow of intake air occurs, for example, to about 110 ° CA after intake bottom dead center (ABDC). If the closing timing of the intake valve 11 is further retarded than the latest timing Tx, the effective compression ratio of the engine is extremely lowered and the output becomes insufficient, so the latest timing Tx is set as an amount that can be retarded to the maximum.

In the specific operation region R, first, the closing timing of the intake valve 11 is set to, for example, about 35 ± 5 ° CA after intake bottom dead center (ABDC) as a timing at which backflow of intake air hardly occurs. Therefore, the first determination in step S41 is no, and the process proceeds to step S43 where control is executed to retard (delay) the closing timing of the intake valve 11. Specifically, by driving the VVT15 in the direction in which the operation timing of the intake valve 11 is retarded, the closing timing of the intake valve 11 is retarded by a predetermined amount from the current setting value, and the effective compression ratio of the engine is lowered.

The delay in the closing timing of the intake valve 11 is performed in stages by dividing the closing timing into a plurality of times, similarly to the case of the enrichment of the air-fuel ratio in step S42. That is, the closing timing of the intake valve 11 is first retarded by a predetermined amount, and if the preignition is avoided in this state, the retard angle is prohibited from being further retarded, while if the preignition cannot be avoided, the retard angle is further increased.

In the present embodiment, when the closing timing of the intake valve 11 is retarded in stages as described above, the retard angle amount is set for each stage so that the effective compression ratio is gradually lowered by a certain interval with the retard angle. Therefore, as the closing timing before the retardation angle is set approaches the intake bottom dead center, the retardation angle amount at each time becomes larger, and as the retardation angle advances, the retardation angle amount at each time becomes smaller.

The reason why the retardation angle is controlled in the above manner will be described with reference to fig. 13. Fig. 13 is a diagram showing a relationship between the retard angle amount of the closing timing (IVC) of the intake valve 11 and the effective compression ratio in the engine having the geometric compression ratio 14. As can be seen from this graph, the rate of decrease in the effective compression ratio gradually increases as the slope of the graph increases as the closing timing of the intake valve 11 becomes farther from intake Bottom Dead Center (BDC) (toward the right side of the lateral axis). Therefore, if the effective compression ratio is to be lowered by a certain amount all the time, the more the current closing timing of the intake valve 11 is retarded with respect to intake bottom dead center, the more the retardation angle amount from that point on needs to be reduced, and conversely, the closer the current closing timing of the intake valve 11 is to intake bottom dead center, the more the retardation angle amount needs to be increased.

Fig. 14 is a diagram showing how the delay angle amount (vertical axis) of the closing timing of the intake valve 11 required to decrease the effective compression ratio by 0.5 changes in the range of the delay angle amount 30 ° CA or more, in accordance with the value of the current delay angle amount (horizontal axis). According to this figure, for example, if the current retard angle amount is 30 ° CA, the effective compression ratio cannot be lowered by 0.5 if the delay angle amount is further about 10 ° CA from this start, whereas if the current retard angle amount is 40 ° CA, the effective compression ratio can be lowered by 0.5 if the delay angle amount is about 8 ° CA from this start. Thus, the larger the current retard angle amount is, the smaller the retard angle amount of the closing timing of the intake valve 11 required to lower the effective compression ratio by a certain amount is.

Therefore, when the closing timing of the intake valve 11 is retarded with respect to the intake bottom dead center in step S43, the retard amount gradually decreases as the closing timing before the retard becomes farther from the intake bottom dead center, and the closing timing of the intake valve 11 is retarded by dividing the range up to the latest timing Tx a plurality of times, thereby gradually decreasing the effective compression ratio by a fixed interval in stages.

After the closing timing of the intake valve 11 is retarded to the latest timing Tx in step S43, if the pre-ignition continues, no determination is made in step S41, and therefore control is executed to divide the fuel injection timing and inject a part of the fuel in the compression stroke in the next step S44. That is, as shown in fig. 15 (a), all the fuel should be injected in the intake stroke (F in the drawing) as it is, but as shown in fig. 15 (b), the injection timing of some of the fuel is delayed to the later stage of the compression stroke, and the fuel is divided into injection in the intake stroke and the compression stroke (F1 and F2 in the drawing).

As described above, in the preignition avoidance control, the enrichment of the air-fuel ratio (step S42), the delay of the closing timing of the intake valve 11 (step S43), and the delay of the fuel injection timing (step S44) are performed in this order with priority.

When any of the above-described controls of steps S42, S43, S44 is started, subsequently, "1" indicating that the pre-ignition avoidance control is being executed is input to the control execution flag FF (whose default value is 0) for recording the execution state of the control (step S45), and the flow returns to the main flow of fig. 10.

Fig. 16 is a time chart showing how the air-fuel ratio (a/F), the closing timing (IVC) of the intake valve 11, and the fuel injection timing change with the passage of time, assuming that the preignition avoidance cannot be achieved without executing all the controls of steps S42, S43, and S44 in the preignition avoidance control. As can be understood from this figure, in the preignition avoidance control, first, the control of enriching the air-fuel ratio in stages is preferentially executed, and at this time, if the preignition cannot be avoided even after the enrichment to the maximum (up to a/F11), the closing timing (IVC) of the intake valve 11 is retarded in stages, and at this time, if the preignition cannot be avoided even after the maximum retardation, the fuel injection timing is retarded (a part of the fuel is injected in the compression stroke).

Referring back to fig. 10 again, the control operation when the determination at step S20 is no will be described. If the pre-ignition is not sufficiently suppressed or if the pre-ignition does not occur from the beginning as a result of the pre-ignition avoidance control (S21), the abnormal combustion flag Fabnrm ≠ 1, and the determination at the above-described step S20 is "no". Then, in the next step S23, it is determined whether or not the abnormal combustion flag Fabnrm is 2 and knocking has occurred.

When it is determined as yes in the above step S23 and it is confirmed that knocking has occurred, control is executed to retard (retard) the ignition timing until the knocking is sufficiently suppressed (step S24), and control is executed to input a flag FF "2" to the above control execution (step S25) in order to record that the control is being executed.

If knocking is sufficiently suppressed by the retard of the ignition timing or if knocking has not occurred from the beginning, the determination at step S23 is no. That is, since step S20 and step S23 are both no, the abnormal combustion flag Fabnrm is 0, and both preignition and knocking do not occur, and the combustion state is normal. Then, in the next step S26, it is determined whether the control execution flag FF is "1", that is, whether the above-described pre-ignition avoidance control (S21) is being executed.

Assuming that the current combustion state becomes normal as a result of the execution of the pre-ignition avoidance control, the flag FF becomes 1, and thus it is determined as yes in the step S26. Then, in the next step S27, a return control for canceling the above-described pre-ignition avoidance control and returning to the normal operation is executed.

Fig. 12 shows the details of the recovery control in step S27. After the recovery control is started, it is first determined whether or not the control (step S44 in fig. 11) for delaying the injection timing of a part of the fuel to the late stage of the compression stroke is being executed (step S50), and when it is determined that "yes" and the late angle of the fuel injection timing (compression stroke injection) is being executed, the control for recovering the injection timing of the part of the fuel to the normal injection timing, that is, the intake stroke is executed (step S53).

As described above, if the injection timing of the fuel is recovered to the regular timing (in the intake stroke) and then the pre-ignition does not occur, or if the retard of the fuel injection timing is not performed from the beginning, the determination in step S50 is no. Then, the process proceeds to step S51, and control is executed to determine whether the closing timing of the intake valve 11 is set at a timing that is delayed from the original setting timing.

If the closing timing of the intake valve 11 has been retarded in the above-described step S43 of fig. 11, it is determined as yes in the above-described step S51. Then, the process proceeds to step S54, and control is executed to return the closing timing of the intake valve 11 to the advance (advance) side to increase the effective compression ratio.

The advancing of the closing timing of the intake valve 11 is performed in stages by dividing the closing timing into a plurality of times, as in step S43 of fig. 11. According to the map of fig. 14, the advance amount at each time is smaller as the closing timing before the advance is farther from intake bottom dead center and larger as it is closer to intake bottom dead center. Such advancement in stages is continued until the closing timing of the intake valve 11 reaches a normal timing (a timing at which backflow of intake air hardly occurs; for example, about ABDC35 ± 5 °), whereby the effective compression ratio is gradually increased at regular intervals and is returned to a value close to the geometric compression ratio.

As described above, if the preignition does not occur after the closing timing of the intake valve 11 is returned to the normal timing, or the closing timing of the intake valve 11 is not delayed from the initial start, the determination at step S51 is no. Then, the process proceeds to step S52, and control is executed to determine whether the air-fuel ratio in the cylinder is richer than a normal value (the stoichiometric air-fuel ratio or its vicinity). Next, if it is determined "yes" here and it is confirmed that the air-fuel ratio is rich, control is performed to return the air-fuel ratio to the lean side (the side close to the normal value) (step S55).

The air-fuel ratio is made lean a plurality of times and is made stepwise, as in step S42 of fig. 11. For example, the air-fuel ratio in the cylinder is made lean in stages so that a/F is 11 → 12.5 → 14.7, and is returned to a normal value.

After the control at step S55 is completed and the air-fuel ratio is returned to a normal value, the determination at step S52 is no. Then, "0" is input to the control execution flag FF (step S56), and the flow returns to the main flow of fig. 10.

Fig. 17 is a timing chart showing temporal changes in the air-fuel ratio, the fuel injection timing, and the like when the recovery control is performed as described above. What is meant here is: when the preignition avoidance control shown in fig. 16 is performed, that is, when all operations of enriching the air-fuel ratio, retarding the closing timing of the intake valve 11, and retarding the fuel injection timing (a part of the fuel is injected in the compression stroke) are necessary in order to avoid preignition, how each state amount changes by the subsequent recovery control.

As shown in fig. 17, when returning from the preignition avoidance control, the fuel injection timing is first retarded to return to the normal timing (during the intake stroke), and if no preignition occurs thereafter, control is executed to advance the closing timing (IVC) of the intake valve 11 in stages toward the normal timing, and if no preignition occurs as such, control is executed to make the air-fuel ratio in stages lean toward the normal value.

Referring back to fig. 10 again, the control operation when the determination at step S26 is no will be described. As a result of the above-described return control, if the pre-ignition avoidance control (S21) is completely released, the air-fuel ratio, the closing timing of the intake valve 11, and the fuel injection timing are all returned to the normal values, or if the pre-ignition avoidance control is not performed from the beginning, the control execution flag FF ≠ 1 is set, and the determination in the above-described step S26 is "no". Then, in the next step S29, it is determined whether or not the control execution flag is FF 2, that is, whether or not the control of retarding the ignition timing to avoid knocking is executed.

If it is determined as yes in the above step S29 and it is confirmed that the ignition timing has been retarded, the ignition timing after the retardation is advanced to return to the normal ignition timing (in the specific operation region R, for example, around ATDC5 °) (step S30), and control is performed in which "0" is input to the above-described control execution flag FF (step S31).

As described above, if knocking does not occur after the ignition timing returns to the normal timing, or if the ignition timing is not retarded from the first time, it is determined as no in step S29 and the normal operation is continued (step S32).

(5) Effect of action

In the spark ignition engine of the present embodiment, in the specific operating region R set in the low rotation speed/high load region of the engine, the flame is detected by the ion current sensor 34, and the presence or absence of preignition is determined based on the measured time (the time of occurrence of flame), while even if preignition is not detected by the detection using the ion current sensor 34, preignition is detected by the vibration sensor 33. According to this structure, the early ignition, which is a phenomenon in which the mixed gas is ignited early, can be detected with high accuracy, separately from the knocking.

Specifically, when detecting preignition using the vibration sensor 33, it is first determined whether or not the maximum vibration intensity Vmax1 obtained from the vibration sensor 33 is equal to or greater than the threshold X (S6), and if so, the ignition timing of the spark plug 16 is further changed from the ignition timing at the normal time that is on a slightly retarded side (e.g., around ATDC5 °) of the compression top dead center to the retarded side. Then, it is determined whether or not the maximum vibration intensity obtained after the retardation of the ignition timing (maximum vibration intensity after ignition retardation) Vmax2 is greater than the maximum vibration intensity before the retardation (maximum vibration intensity before ignition retardation) Vmax1(S10), and if Vmax2 is greater than Vmax1, it is determined that pre-ignition has occurred. By performing such a procedure, there is an advantage that even a relatively early-stage preignition (for example, a light preignition as shown in a waveform J1 of fig. 4 or an preignition close thereto) which does not progress to a certain extent can be reliably detected while being distinguished from knocking.

For example, if only the maximum vibration intensity Vmax is simply compared with the reference value, it is difficult to determine whether the pre-ignition is the relatively early stage or the knocking. In view of such a problem, in the above configuration, when the maximum vibration intensity Vmax is confirmed to be equal to or greater than a predetermined threshold, the ignition timing is intentionally retarded, and when an increase in the maximum vibration intensity Vmax is confirmed at the preceding and following ignition timings, it is determined that preignition has occurred. That is, since the late angle of the ignition timing is effective only for suppressing knocking (ineffective for suppressing preignition), by utilizing this phenomenon, it is possible to accurately determine whether preignition or knocking is caused by examining the change in the maximum vibration intensity Vmax after the late angle of the ignition timing.

Therefore, according to the above configuration, even if the vibration sensor 33 is used, the occurrence of preignition, which is a phenomenon in which the mixed gas is ignited early enough, can be reliably detected, separately from knocking, and if a failure such as disconnection occurs in the ion current sensor 34 and the detection accuracy is insufficient, the occurrence of preignition is not missed. When the pre-ignition is detected, it is possible to reliably prevent an engine failure (for example, damage to the piston 5) due to the continued pre-ignition by taking necessary measures for avoiding this phenomenon (i.e., control of the enrichment of the air-fuel ratio, reduction of the effective compression ratio, and the like).

In the above embodiment, even when the maximum vibration intensity Vmax2 after the ignition delay is equal to or less than the maximum vibration intensity Vmax1 before the ignition delay, it can be determined that the preignition has occurred if the measured time of Vmax2 is earlier than the measured time of Vmax 1. That is, when the preignition occurs, even if the ignition timing is retarded, the phenomenon that the maximum vibration intensity Vmax increases and the measurement timing thereof is advanced should be observed, but depending on the case, only one of the phenomena may be observed, and therefore, it is determined that the preignition occurs only by confirming the increase of the maximum vibration intensity Vmax or the advancement of the measurement timing thereof. Therefore, the detection precision of the pre-ignition can be improved.

In the above-described embodiment, when the occurrence of pre-ignition is confirmed based on the detection values of the ion current sensor 34 and the vibration sensor 33 in the specific operation region R (yes in each of S3, S10, and S11), pre-ignition avoidance control (S21) is executed as control for avoiding this. In the pre-ignition avoidance control, first, control is performed to increase the injection amount of fuel injected from the injector 18 to enrich the air-fuel ratio in the cylinder (S42), and if pre-ignition is detected even after the control, control is performed to retard the closing timing of the intake valve 11 to lower the effective compression ratio of the engine (S43), and when pre-ignition is detected even as it is, control is performed to retard the injection timing of a part of the fuel to the late stage of the compression stroke (S44). According to this structure, there is an advantage that not only the emission performance can be maintained as good as possible, but also the occurrence of pre-ignition can be effectively suppressed.

That is, in the above-described embodiment, in the preignition avoidance control for avoiding preignition, the control for enriching the air-fuel ratio is performed first and the control for delaying the fuel injection timing (injecting a part of the fuel in the compression stroke) is performed last, so that it is possible to effectively control the occurrence of preignition while avoiding the occurrence of smoke and deterioration of emission performance as much as possible.

In a spark ignition engine, even if the air-fuel ratio is made rich or the fuel injection timing is retarded, the in-cylinder temperature can be reduced to suppress pre-ignition, but the retarded timing of the fuel injection (compression stroke injection) may cause smoke generation due to the remaining of a large amount of unburned carbon components, and therefore if the injection timing is retarded immediately, smoke may be generated frequently. In contrast, in the above embodiment, the air-fuel ratio is first made rich to reduce the in-cylinder temperature when the pre-ignition occurs, and the retarding of the injection timing is executed only when the pre-ignition cannot be avoided yet, so that there is an advantage that the generation of smoke can be avoided as much as possible and the emission performance can be maintained as good as possible.

Further, since the control (S43) is executed to retard the closing timing of the intake valve 11 and lower the effective compression ratio as the control of the retarded angle (S44) whose priority is lower than the fuel enrichment (S42) but higher than the fuel injection timing, there is an advantage that the retarded angle at the fuel injection timing can be made less frequent, and the deterioration of the emission performance due to the generation of smoke can be more effectively prevented.

That is, after the air-fuel ratio is made rich, control is executed to decrease the effective compression ratio of the engine to achieve a decrease in the in-cylinder pressure, and the fuel injection timing is retarded only when it is not possible to avoid pre-ignition as such, so that pre-ignition can be avoided with a higher probability without retarding the fuel injection timing. This makes it possible to reduce the frequency of late ignition in the fuel injection timing, to avoid smoke generation as much as possible, and to suppress pre-ignition.

In the above configuration, when the transition is made to the pre-ignition avoidance control, the control for making the air-fuel ratio rich is executed first, and when the pre-ignition cannot be avoided as such, the control for delaying the closing timing of the intake valve 11 to lower the effective compression ratio is executed. However, the control for lowering the effective compression ratio causes not only a decrease in the output of the engine but also poor responsiveness of the control. That is, particularly when VVT15 is a hydraulic variable mechanism, a relatively long response delay occurs in changing the operation timing of the intake valve 11 by driving VVT15, and therefore control in which the closing timing of the intake valve 11 is delayed to lower the effective compression ratio is inferior in control responsiveness to control in which the air-fuel ratio is made rich by increasing the injection amount injected from the injector 18.

Therefore, in the above embodiment, at the time of the pre-ignition avoidance control, the enrichment of the air-fuel ratio is performed first, and then the reduction of the effective compression ratio is performed. In this way, by setting the reduction of the effective compression ratio to the rear, it is possible to avoid the reduction of the engine output due to the reduction of the compression ratio as much as possible, and by making the air-fuel ratio rich with excellent responsiveness the highest priority, there is an advantage that the suppression thereof can be achieved more quickly after the occurrence of the preignition.

In the above embodiment, the rich state of the air-fuel ratio is performed in stages during the pre-ignition avoidance control, and if the pre-ignition is detected even in a state where the pre-ignition is rich to the maximum (the a/F to the threshold value is 11), the control for lowering the effective compression ratio is performed (see fig. 16). In this way, when the air-fuel ratio is made rich in stages, for example, when the degree of pre-ignition is small and the pre-ignition can be avoided by making the air-fuel ratio slightly rich, the deterioration of fuel efficiency and the like due to the rich air-fuel ratio can be suppressed to the minimum without making the air-fuel ratio excessively rich. Further, when the pre-ignition cannot be avoided even if the air-fuel ratio is made maximally rich, the pre-ignition can be suppressed by lowering the effective compression ratio and further by retarding the fuel injection timing, so that the pre-ignition that has progressed to a certain extent can be reliably avoided while preventing the air-fuel ratio from being made excessively rich.

In the above embodiment, the control for retarding the closing timing of the intake valve 11 to lower the effective compression ratio is also performed in stages, and if pre-ignition is detected even in a state where the effective compression ratio is lowered to the maximum (up to the compression ratio corresponding to the latest timing Tx), the control for retarding the fuel injection timing is executed. According to this configuration, there is an advantage that the engine output can be prevented from being greatly reduced due to an excessive reduction in the effective compression ratio, and the preignition can be avoided more reliably.

In particular, in the above embodiment, when the closing timing of the intake valve 11 is retarded in stages, as shown in fig. 14, the retard angle amount from this point is set larger as the current closing timing approaches intake bottom dead center, so the effective compression ratio can be lowered at regular intervals each time the closing timing of the intake valve 11 is retarded. Therefore, it is possible to appropriately avoid a situation in which the engine output is rapidly reduced by 1-time retarded angle, or the compression ratio is only slightly reduced and the effect on the preignition is hardly obtained, and there is an advantage that the occurrence of the preignition can be suppressed more effectively.

In the above embodiment, the closing timing of the intake valve 11 at the normal time (when the pre-ignition does not occur) is set to the timing on the retard side of the intake bottom dead center and the reverse flow of the intake air is hardly caused (about ABDC35 ± 5 ° in the specific operating region R), and on the other hand, when the effective compression ratio is lowered by the pre-ignition avoidance control, the VVT15 is driven to further retard the closing timing of the intake valve 11 with respect to the intake bottom dead center, so that there is an advantage that the engine output at the normal time can be sufficiently ensured and the effective compression ratio can be efficiently lowered when necessary.

For example, if the effective compression ratio is simply lowered, the effective compression ratio can be lowered even if the closing timing of the intake valve 11 is advanced to the advanced side of intake bottom dead center. However, in the case where the closing timing of the intake valve 11 at the normal time is on the retarded side of intake bottom dead center, if the effective compression ratio is to be lowered by changing the closing timing from the retarded side to the advanced side of intake bottom dead center, the operation timing of the intake valve 11 must be changed greatly, and the controlled variable of the VVT15 increases, which causes a problem of deterioration in the responsiveness of the control. In order to avoid this, it is conceivable to set the closing timing of the intake valve 11 at the normal time to a timing substantially matching the intake bottom dead center or to a timing advanced from this timing, but this makes it impossible to fully utilize the intake inertia, resulting in a decrease in the engine output.

From this viewpoint, it is also advantageous in the above-described embodiment that the closing timing of the intake valve 11 in the normal state is set on the retard side of the intake bottom dead center, and the closing timing of the intake valve 11 is retarded with respect to the normal timing when the effective compression ratio is lowered, in that the engine output in the normal state is sufficiently ensured and the effective compression ratio can be efficiently lowered when necessary.

In the above embodiment, when returning from the pre-ignition avoidance control to the normal operation, the control of canceling the late angle of the fuel injection timing (injecting a part of the fuel in the compression stroke) is executed with the highest priority, and the injection timing of the part of the fuel delayed to the latter stage of the compression stroke is returned to the intake stroke.

For example, in the preignition avoidance control, when the enrichment of the air-fuel ratio, the decrease of the effective compression ratio (the late angle of the closing timing of the intake valve 11), the late angle of the fuel injection timing, and the like are all necessary, when returning from this state to the normal operation, as shown in fig. 17, the late angle of the fuel injection timing is first released to return the injection timing to the intake stroke, and when preignition is not detected thereafter, the closing timing of the intake valve 11 is returned to the advanced angle side to increase the effective compression ratio, and even if preignition is not detected as such, the air-fuel ratio is returned to the lean side. According to such a configuration, when the pre-ignition is avoided, the delay of the fuel injection timing is first canceled to remove the possibility of smoke generation, thereby making it possible to minimize the time during which the emission performance deteriorates.

Further, if no preignition is detected subsequently, as control to be given priority next, the closing timing of the intake valve 11 is advanced to increase the effective compression ratio, whereby a decrease in engine output due to a decrease in the compression ratio can be canceled early. If even this is the case, the pre-ignition is not detected, and the air-fuel ratio is finally returned to the lean side, whereby it is possible to ensure that no pre-ignition occurs and to appropriately return to the normal operation.

(6) Modifications and the like

In the above embodiment, when the air-fuel ratio is made rich in the pre-ignition avoidance control, for example, a/F is 14.7 → 12.5 → 11, the enrichment is performed in multiple steps and multiple times, but the air-fuel ratio may be made rich in multiple steps and multiple times. Conversely, if the number of times of enrichment is set to only 1 and then pre-ignition cannot be avoided, it is also possible to immediately transition to control in which the closing timing of the intake valve 11 is retarded to lower the effective compression ratio.

In the above embodiment, when the closing timing of the intake valve 11 is retarded to lower the effective compression ratio in the pre-ignition avoidance control, the closing timing of the intake valve 11 is retarded in stages by dividing it into a plurality of times, but the number of times of retard angle can be set as appropriate depending on the characteristics of the engine and the like.

Further, if it is desired not to reduce the output as much as possible in terms of the engine characteristics, the delay of the closing timing of the intake valve 11 may be set to only 1 time. In this case, however, the more retarded the closing timing of the intake valve 11 before the retardation angle is set with respect to intake bottom dead center, the smaller the retardation angle amount from that time should be. That is, in the specific operating region R where pre-ignition is likely to occur, the closing timing of the intake valve 11 at the normal time has a certain width, for example, about 35 ± 5 ° after passing through intake bottom dead center (ABDC), so if the closing timing of the intake valve 11 before the start of the retard is, for example, ABDC40 °, the retard amount at the time of retarding the closing timing is set smaller than that at the time of ABDC30 °. With this, the reduction width of the effective compression ratio can be made constant regardless of the closing timing of the intake valve 11 at the normal time.

In the above embodiment, for example, as shown in fig. 15 (a), the normal fuel injection timing in which no preignition occurs is set to 1 time in the intake stroke (i.e., all fuel is injected 1 time in the intake stroke), but the normal fuel injection timing may be in the intake stroke, or the fuel may be injected in multiple portions in the intake stroke.

In the above-described embodiment, when the pre-ignition is detected, the fuel enrichment of the air-fuel ratio and the reduction of the effective compression ratio are performed in sequence, and when the pre-ignition cannot be avoided, the injection timing of a part of the fuel to be injected is retarded to the later stage of the compression stroke (fig. 15 (b)) in a stepwise manner, but for example, as shown in fig. 18 (a) to (c), the second injection F2 (hereinafter, referred to as the latter stage injection) retarded in the compression stroke may be retarded in a stepwise manner from the middle stage to the later stage of the compression stroke. That is, first, the timing of the latter injection F2 is retarded to the middle stage of the compression stroke ((b) of fig. 18), and if the pre-ignition cannot be avoided in this state, the timing of the latter injection F2 is further retarded to be set in the latter stage of the compression stroke ((c) of fig. 18). Thus, if the pre-ignition can be sufficiently avoided by delaying the latter injection F2 to the middle stage of the compression stroke, the injection timing does not need to be delayed at once to the latter stage of the compression stroke where the possibility of smoke generation is high, and deterioration of the emission performance can be more effectively prevented.

Conversely, it is also conceivable that the pre-ignition cannot be avoided even if the late injection F2 is delayed to the late stage of the compression stroke, based on the characteristics of the engine, or the like. Therefore, in this case, for example, simultaneously with or after the control of delaying the above-described post-injection F2 to the late stage of the compression stroke, the a/F at which the air-fuel ratio is changed to the threshold value may be made richer (for example, to about 10). Thereby, although the possibility of smoke generation becomes higher temporarily, even in the case where the preignition has progressed considerably, it can be reliably avoided.

In the above embodiment, the control is performed to divide the fuel injection and retard a part thereof (the latter injection F2) into the compression stroke, and if the result of this control is that the pre-ignition is avoided, the injection timing of the part of the fuel retarded into the compression stroke is immediately returned to the normal timing (into the intake stroke), but the injection timing after the pre-ignition is avoided may be returned to at least the advance angle side (the side close to the intake stroke), or may be advanced to the normal timing in stages.

In the above embodiment, the in-cylinder temperature and the in-cylinder pressure are reduced by preferentially executing the enrichment of the air-fuel ratio (S42), the reduction of the effective compression ratio (S43), and the compression stroke injection of a part of the fuel (S44) in this order as the pre-ignition avoidance control, but controls other than the three types of controls may be adopted as long as at least one of the in-cylinder temperature and the in-cylinder pressure can be reduced. For example, a cooling device for cooling intake air may be provided in the middle of the intake passage 20, and the intake air cooled by this cooling device may be introduced into the combustion chamber 6.

In the above embodiment, the ion current sensor 34 is provided separately from the ignition plug 16, and whether or not pre-ignition has occurred is determined by detecting the timing of flame generation using the ion current sensor 34, but the ignition plug 16 may also be used as the ion current sensor 34 as long as a bias voltage for detecting ion current can be applied to the center electrode (ignition plug electrode) of the ignition plug 16. In this way, the structure can be simplified, and the cost associated with the ion current sensor 34 can be reduced.

However, if the spark plug 16 is used as the ion current sensor 34 as described above, the spark plug electrode cannot be biased at the moment when the spark is discharged from the spark plug 16 (i.e., at the moment when the high voltage for discharge is applied to the spark plug electrode), and the ion current cannot be detected, so that the accuracy of detecting the preignition by the single ion current sensor 34 is lowered. However, in the configuration for detecting the preignition using the vibration sensor 33 according to the above embodiment, since the reduction in the detection accuracy can be compensated for by the vibration sensor 33, the reduction in the detection accuracy of the preignition can be avoided, and the configuration can be simplified and the component cost can be reduced.

In the above embodiment, the occurrence of the preignition is detected using both the ion current sensor 34 and the vibration sensor 33, but the detection of the preignition by the ion current sensor 34 may be omitted. In this way, the occurrence of pre-ignition can be detected by the single vibration sensor 33, and the structure and control can be further simplified, thereby further reducing the component cost.

In the above-described embodiment, the vibration of the engine body 1 is detected by the vibration sensor 33, and whether preignition or knocking has occurred is determined based on how the magnitude of the maximum vibration intensity Vmax specified from the detected value and the measurement timing thereof change as the ignition timing is retarded.

Specifically, the detection of pre-ignition and knocking using the in-cylinder pressure sensor is performed in the following manner. First, when the operating state of the engine is in the specific operating region R, the maximum value of the in-cylinder pressure is specified from the waveform of the in-cylinder pressure input from the in-cylinder pressure sensor (see fig. 5, for example), and it is determined whether or not the maximum value is equal to or greater than a predetermined threshold value. If the value is above the threshold value, the ignition timing is retarded, and then the maximum value of the in-cylinder pressure is acquired again. Next, the maximum value of the in-cylinder pressure obtained after the ignition timing is retarded is set as the maximum in-cylinder pressure after the ignition delay (corresponding to the maximum detection value after the ignition delay according to the present invention), the maximum value of the in-cylinder pressure obtained before the ignition timing is retarded is set as the maximum in-cylinder pressure before the ignition delay (corresponding to the maximum detection value before the ignition delay according to the present invention), and it is determined whether or not the maximum in-cylinder pressure after the ignition delay is larger than the maximum in-cylinder pressure before the ignition delay, and if so, it is determined that the preignition has occurred.

Further, as in the above-described embodiment, even if the maximum cylinder internal pressure after the ignition delay is equal to or less than the maximum cylinder internal pressure before the ignition delay, it can be determined that the pre-ignition is generated as long as the measurement timing of the maximum cylinder internal pressure after the ignition delay is earlier than the measurement timing of the maximum cylinder internal pressure before the ignition delay. On the other hand, if the measured timing after the ignition delay is not earlier than before the ignition delay, knocking is determined.

(7) Summary of the invention

Finally, the structure of the present invention disclosed based on the above-described embodiments and the effects thereof will be summarized.

The method of the present invention is an abnormal combustion detection method for a spark ignition engine provided with a vibration sensor for detecting vibration of the engine or an in-cylinder pressure sensor for detecting an in-cylinder pressure, the engine setting an ignition timing of an ignition plug at a normal time in a low rotation speed and high load region and in which abnormal combustion does not occur on a retard angle side of a compression top dead center, the method including the steps of: determining whether or not a maximum value of the vibration intensity obtained from the vibration sensor or a maximum value of the in-cylinder pressure obtained from the in-cylinder pressure sensor in a low rotation speed and high load region is equal to or greater than a predetermined threshold value; changing the ignition timing of the spark plug from the ignition timing of the normal time on the retard side of the compression top dead center to a further retard side when the maximum value of the vibration intensity or the maximum value of the in-cylinder pressure is equal to or greater than the predetermined threshold value; when the maximum value of the vibration intensity obtained from the vibration sensor after the ignition timing is changed to the retard side or the maximum value of the in-cylinder pressure obtained from the in-cylinder pressure sensor is set to the maximum detection value after the ignition delay and the maximum value of the vibration intensity obtained before the ignition timing is changed to the retard side or the maximum value of the in-cylinder pressure is set to the maximum detection value before the ignition delay, it is determined that preignition of the mixture is caused if the maximum detection value after the ignition delay is larger than the maximum detection value before the ignition delay.

An engine according to the present invention is a spark ignition type engine including a vibration sensor for detecting vibration of the engine or an in-cylinder pressure sensor for detecting an in-cylinder pressure, the engine setting an ignition timing of an ignition plug at a normal time in a low rotation speed and high load region and in which abnormal combustion does not occur on a retard side of a compression top dead center, the engine including: a control unit that controls a timing of spark discharge of the spark plug and receives information of a vibration intensity detected by the vibration sensor or information of an in-cylinder pressure detected by the in-cylinder pressure sensor; wherein the control means executes control to change the ignition timing of the spark plug from the ignition timing of the normal time on the retard angle side located at a compression top dead center to the retard angle side further when a maximum value of the vibration intensity acquired from the vibration sensor or a maximum value of the in-cylinder pressure acquired from the in-cylinder pressure sensor in a low rotation speed and high load region is equal to or more than a predetermined threshold value, and if the maximum value of the vibration intensity acquired from the vibration sensor or the maximum value of the in-cylinder pressure acquired from the in-cylinder pressure sensor after the ignition timing is changed to the retard angle side is a maximum detection value after an ignition delay and the maximum value of the vibration intensity acquired before the ignition timing is changed to the retard angle side or the maximum value of the in-cylinder pressure is a maximum detection value before the ignition delay, if the maximum detection value after the ignition delay is larger than the maximum detection value before the ignition delay, it is determined that preignition in which the mixture is ignited prematurely occurs.

According to the above-described invention, the maximum value of the vibration intensity or the maximum value of the in-cylinder pressure is acquired by using the vibration sensor or the in-cylinder pressure sensor, the ignition timing is retarded when the maximum value is equal to or larger than a predetermined threshold value, and the presence or absence of the occurrence of the preignition is determined based on whether or not the maximum value of the vibration intensity or the in-cylinder pressure (the maximum detection value after the ignition delay) after the ignition timing is changed to the retard side (retarded angle) is larger than the maximum value (the maximum detection value before the ignition delay) before the ignition timing is changed to the retard side (retarded angle), so that there is an advantage that even in the case of the relatively early-stage preignition which is not so severe, the preignition can be distinguished from the knocking and the preignition can be reliably detected.

For example, if the vibration intensity or the in-cylinder pressure is simply compared with the reference value, it is difficult to determine whether the pre-ignition is a relatively early stage or not, particularly when the pre-ignition is a relatively early stage. In view of such a problem, in the configuration of the present invention described above, when the maximum detection value (the maximum value of the vibration intensity or the in-cylinder pressure) of the vibration sensor or the in-cylinder pressure sensor is equal to or greater than a predetermined threshold value, the ignition timing is intentionally retarded, and when an increase in the maximum detection value is confirmed before or after the retardation, it is determined that preignition has occurred. That is, since the retarding of the ignition timing is effective only for knocking suppression (ineffective for suppressing preignition), by utilizing this effect, the change in the maximum detection value after the retarding of the ignition timing is examined, and it is possible to accurately determine whether the ignition is preignition or knocking.

In the detection method of the present invention, it is preferable that, even when the maximum detection value after the ignition delay is equal to or less than the maximum detection value before the ignition delay, if the measurement timing of the maximum detection value after the ignition delay is earlier than the measurement timing of the maximum detection value before the ignition delay, it is determined that preignition has occurred.

In the engine according to the present invention, it is preferable that the control means determines that preignition has occurred if the timing of measurement of the maximum detection value after ignition delay is earlier than the timing of measurement of the maximum detection value before ignition delay even when the maximum detection value after ignition delay is equal to or less than the maximum detection value before ignition delay.

Since the pre-ignition progresses gradually regardless of the retarded ignition timing, the ignition timing advances with the passage of time once the pre-ignition occurs. In the above-described aspect, even if an increase in the maximum detection value of the vibration sensor or the in-cylinder pressure sensor is not observed, the preignition is determined as long as the measurement timing thereof is advanced. This can further improve the detection accuracy of preignition.

In the detection method of the present invention, it is preferable that the engine is provided with an ion current sensor for detecting a flame generated by combustion of the air-fuel mixture, whether or not preignition occurs is determined based on a flame measurement timing of the ion current sensor, and even when it is determined that no preignition has occurred as a result of the determination, it is determined that preignition has occurred if it is determined that a maximum detection value after the ignition delay detected by the vibration sensor or the in-cylinder pressure sensor is larger than a maximum detection value before the ignition delay.

The engine of the present invention preferably further comprises: an ion current sensor that detects a flame generated based on combustion of the mixed gas; wherein the control means determines whether or not there is a preignition based on a flame measurement timing of the ion current sensor, and determines that the preignition has occurred if it is determined that a maximum detection value after the ignition delay detected by the vibration sensor or the in-cylinder pressure sensor is greater than a maximum detection value before the ignition delay even if the determination result indicates that the preignition has not occurred.

When a dual detection system is constructed using the ion current sensor and the vibration sensor or the in-cylinder pressure sensor in combination as in these proposals, even if a failure such as disconnection occurs in the ion current sensor and the detection accuracy is insufficient, the vibration sensor or the in-cylinder pressure sensor can be used to detect preignition, and the detection accuracy can be further improved.

In the engine according to the present invention, preferably, the control means executes a predetermined control for decreasing at least one of the in-cylinder temperature and the in-cylinder pressure when it is determined that the pre-ignition has occurred.

According to this aspect, there is an advantage in that, when the pre-ignition occurs, the occurrence of the pre-ignition can be effectively suppressed by reducing at least one of the in-cylinder temperature and the in-cylinder pressure.

Further, as the control for mainly lowering the in-cylinder temperature, for example, a control for enriching the air-fuel ratio and a control for injecting at least a part of the fuel in the compression stroke may be considered. Further, as the control for mainly reducing the in-cylinder pressure, for example, a control for changing the closing timing of the intake valve to reduce the effective compression ratio may be considered. By executing any one of these controls or executing two or more controls at the same time, at least one of the in-cylinder temperature and the in-cylinder pressure can be reduced.

Claims

1. An abnormal combustion detection method for a spark ignition engine provided with a vibration sensor for detecting vibration of the engine or an in-cylinder pressure sensor for detecting in-cylinder pressure, wherein the ignition timing of an ignition plug at a normal time in a low rotation speed and high load region and in which abnormal combustion does not occur is set on a retard side of a compression top dead center,

the abnormal combustion detection method is characterized by comprising the steps of:

determining whether or not a maximum value of the vibration intensity obtained from the vibration sensor or a maximum value of the in-cylinder pressure obtained from the in-cylinder pressure sensor in a low rotation speed and high load region is equal to or greater than a predetermined threshold value;

changing the ignition timing of the spark plug from the normal ignition timing on a retard side of compression top dead center to a retard side when the maximum value of the vibration intensity or the maximum value of the in-cylinder pressure is equal to or greater than the predetermined threshold value; wherein,

when the maximum value of the vibration intensity obtained from the vibration sensor after the ignition timing is changed to the retard side or the maximum value of the in-cylinder pressure obtained from the in-cylinder pressure sensor is set as the maximum detection value after the ignition delay and the maximum value of the vibration intensity obtained before the ignition timing is changed to the retard side or the maximum value of the in-cylinder pressure is set as the maximum detection value before the ignition delay, it is determined that preignition in which the mixed gas is ignited by early self occurs if the maximum detection value after the ignition delay is larger than the maximum detection value before the ignition delay.

2. The abnormal combustion detection method of a spark ignition engine according to claim 1, characterized in that:

even when the maximum detection value after the ignition delay is equal to or less than the maximum detection value before the ignition delay, it is determined that preignition has occurred if the measurement timing of the maximum detection value after the ignition delay is earlier than the measurement timing of the maximum detection value before the ignition delay.

3. The abnormal combustion detection method of a spark ignition engine according to claim 1, characterized in that:

the engine is provided with an ion current sensor that detects a flame generated by combustion of a mixed gas, and the presence or absence of pre-ignition is determined based on a flame measurement timing of the ion current sensor, and even when it is determined that no pre-ignition has occurred as a result of the determination, it is determined that pre-ignition has occurred if it is determined that a maximum detection value after the ignition delay detected by the vibration sensor or the in-cylinder pressure sensor is greater than a maximum detection value before the ignition delay.

4. A spark ignition type engine provided with a vibration sensor for detecting vibration of the engine or an in-cylinder pressure sensor for detecting in-cylinder pressure, the engine setting an ignition timing of an ignition plug at a retard angle side of a compression top dead center in a normal time in a low rotation speed and high load region and in which abnormal combustion does not occur, characterized by comprising:

a control unit that controls a timing of spark discharge of the spark plug, and receives information of a vibration intensity detected by the vibration sensor or information of an in-cylinder pressure detected by the in-cylinder pressure sensor; wherein,

the control means executes control of changing the ignition timing of the ignition plug further to a retard angle side from the ignition timing at regular time on the retard angle side of a compression top dead center when a maximum value of the vibration intensity acquired from the vibration sensor or a maximum value of the in-cylinder pressure acquired from the in-cylinder pressure sensor in a low rotation speed and high load region is equal to or more than a predetermined threshold value, and if the maximum detection value after the ignition delay is larger than the maximum detection value before the ignition delay when the maximum value of the vibration intensity acquired from the vibration sensor or the maximum value of the in-cylinder pressure acquired before the ignition timing is changed to the retard angle side is set as a maximum detection value after the ignition delay and the maximum detection value before the ignition delay is set as a maximum detection value before the ignition delay, it is determined that preignition in which the mixture is ignited prematurely occurs.

5. The spark ignition engine of claim 4, characterized in that:

the control unit determines that pre-ignition occurs if the timing of measurement of the maximum detection value after the ignition delay is earlier than the timing of measurement of the maximum detection value before the ignition delay even when the maximum detection value after the ignition delay is equal to or less than the maximum detection value before the ignition delay.

6. The spark ignition engine of claim 4, characterized by further comprising:

an ion current sensor that detects a flame generated based on combustion of the mixed gas; wherein,

the control means determines the presence or absence of pre-ignition based on the flame measurement timing of the ion current sensor, and determines that pre-ignition occurs if it is determined that the maximum detection value after the ignition delay detected by the vibration sensor or the in-cylinder pressure sensor is greater than the maximum detection value before the ignition delay, even if it is determined that no pre-ignition occurs as a result of the determination.

7. The spark ignition engine according to any one of claims 4 to 6, characterized in that:

the control means executes a predetermined control for decreasing at least one of the in-cylinder temperature and the in-cylinder pressure when it is determined that the pre-ignition has occurred.